On Tuesday, October 23, 2012 22:16:18 Lindsay Mathieson wrote: > I don't know if the following is a bug in QtWebKit-2.3 or whether it is > changed behaviour that has triggered a bug in KWebWallet.
Given the backtrace ends up with many cyclic calls into JavaScriptCore I would tend to think it's related more to QtWebKit than KWebWallet. In a similar case with KHTML/KJS, the problem was due to using -ffast-math, which caused a miscompilation of some of the JavaScript internals (which are highly optimized and so sensitive to optimization bugs). -ffast-math shouldn't be a default optimization option (I think even -O3 doesn't pull it in) but make sure it's not in use. Regards, - Michael Pyne
signature.asc
Description: This is a digitally signed message part.
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<
